Presentation is loading. Please wait.

Presentation is loading. Please wait.

數值方法 2008, Applied Mathematics NDHU1 Chaos time series.

Similar presentations


Presentation on theme: "數值方法 2008, Applied Mathematics NDHU1 Chaos time series."— Presentation transcript:

1 數值方法 2008, Applied Mathematics NDHU1 Chaos time series

2 數值方法 2008, Applied Mathematics NDHU2 Matlab codes for time series data Data : Chaos Time Series Eric's Home Page

3 數值方法 2008, Applied Mathematics NDHU3 load henon.dat n=length(henon); x1=henon(1:2:n-1); x2=henon(2:2:n); plot(x1,x2,'.')

4 數值方法 2008, Applied Mathematics NDHU4 load MG17.dat n=length(MG17); plot(1:1:n,MG17)

5 數值方法 2008, Applied Mathematics NDHU5 load MG30.dat n=length(MG30); plot(1:1:n,MG30)

6 數值方法 2008, Applied Mathematics NDHU6 load ikeda.dat; plot(ikeda(:,1),ikeda(:,2),'.');

7 數值方法 2008, Applied Mathematics NDHU7 load laser.dat n=length(laser); plot(1:1:n,laser)

8 數值方法 2008, Applied Mathematics NDHU8 >> load lorenz.dat; >> n=size(lorenz,1); >> plot(1:1:n,lorenz(:,1),'r') >> plot3(lorenz(:,1),lorenz(:,2),lorenz(:,3),'.');

9 數值方法 2008, Applied Mathematics NDHU9

10 10

11 數值方法 2008, Applied Mathematics NDHU11 Nonlinear recursions for Lorenz Series

12 數值方法 2008, Applied Mathematics NDHU12 load MG17.dat n=length(MG17); plot(1:1:n,MG17) Non-linear recursion

13 數值方法 2008, Applied Mathematics NDHU13 MLPotts Learning

14 數值方法 2008, Applied Mathematics NDHU14 Numerical Result

15 數值方法 2008, Applied Mathematics NDHU15 Experiment  Use the first 1000 data to form a training set  Use MLPotts learning to construct nonlinear recursion  Use the obtained MLPotts network to predict data within time frame 1000:1100

16 數值方法 2008, Applied Mathematics NDHU16 MLP learning learn_MLP.m eval_MLP2.m fa2d.m


Download ppt "數值方法 2008, Applied Mathematics NDHU1 Chaos time series."

Similar presentations


Ads by Google